<< Click to Display Table of Contents >> Navigation:  Building blocks of QGeNIe > Inference algorithms > Clustering algorithm

Clustering algorithm is the fastest known exact algorithm for belief updating in Bayesian networks. It was originally proposed by Lauritzen and Spiegelhalter (1988) and improved by several researchers, e.g., Jensen et al. (1990) or Dawid (1992). Our implementation of the clustering algorithm is very efficient and lightning fast. In fact, it is quite possibly the fastest implementation of this algorithm in existence.

The clustering algorithm is the default algorithm used by QGeNIe. The clustering algorithm is QGeNIe's default algorithm and should be sufficient for most applications. Only when networks become very large and complex, the clustering algorithm may not be fast enough. In that case, it is suggested that the user choose an approximate algorithm offered by the program, the EPIS-BN algorithm (Yuan & Druzdzel, 2003).